ChatGPT vs Claude

ChatGPT and Claude cost the same $20 a month at their main paid tier, but they win at different tasks. Claude leads the public leaderboards for writing and coding as of Nov 6, 2026, while ChatGPT makes images, has the more generous free plan and offers cheaper and pricier plans that Claude lacks.

16:16 Claude is ahead of ChatGPT for writing and coding as of Nov 6, 2026: Anthropic's Claude Opus 5.5 ranks 3rd on the Arena text leaderboard against OpenAI's best at 22nd, and 1st on Arena's web development board. ChatGPT is the better pick for making images (Claude does not generate them), for a free plan with unlimited text chats, and for anyone who wants a plan cheaper or more powerful than $20. Both main paid plans cost $20 a month (ChatGPT Plus and Claude Pro). The gaps on the leaderboards are a few points, not a rout, and each product changes month to month, so every verdict below names its source and date.

Learn more: Claude on Ai.vi.ki and Claude vs ChatGPT on the Claude viki.

Quick facts

Free plans	ChatGPT Free: unlimited text chats with limits on tools; Claude Free: limited usage every 5 hours and every week
Entry paid plan	ChatGPT Go $8 a month; Claude has none below Pro
Main paid plan	ChatGPT Plus $20 a month; Claude Pro $20 a month ($17 a month billed annually)
Top plans	ChatGPT Pro $100, $200 or $500 a month; Claude Max $100 and Ultra $200 a month
Flagship models	GPT-6 Astra (OpenAI); Claude Opus 5.5 and Claude Fable 5.1 (Anthropic)
Image generation	ChatGPT yes; Claude no (diagrams and charts only)
Coding agent	Codex (limited on Free); Claude Code (Pro and up)
Ads	Free and Go plans may show ads; Plus and up never do

Sources, as of Nov 6, 2026: ChatGPT pricing, Codex pricing, Claude pricing, Can Claude produce images?, Ads in ChatGPT.

Verdict by task

Task	ChatGPT	Claude	Verdict
Writing	Best OpenAI model 22nd on Arena text (1485)	Opus 5.5 3rd (1509); won Tom's Guide's writing rounds	Claude
Coding	GPT-6 Astra 2nd on Arena WebDev (1773)	Opus 5.5 1st (1810)	Claude, narrowly
Research	Deep research on every plan, Free limited	Research on Pro and up only	ChatGPT on free plans; no reliable quality comparison yet
Images	ChatGPT Images 2.6 on every plan	No image generation	ChatGPT
Voice	Live voice with published hours (3 a day on Plus)	Voice mode in beta on all plans	ChatGPT
Price	$0, $8, $20, $100, $200, $500	$0, $20, $100, $200	Tie at $20; ChatGPT has more choices
Privacy	Training on by default, can be turned off	Trains on chats only if you allow it	Claude
Limits	Unlimited text chats on Free	Free limited every 5 hours and every week	ChatGPT on free plans

Sources for the table: Arena text leaderboard, updated Oct 27, 2026, Arena WebDev leaderboard, updated Oct 27, 2026, Tom's Guide, Sep 24, 2026, ChatGPT pricing, Claude pricing, ChatGPT Voice, Claude voice mode, Data controls in ChatGPT, Is my data used for model training?, all read Nov 6, 2026.

Writing

Claude wins writing on the evidence available as of Nov 6, 2026. On the Arena text leaderboard, where people vote blind between two answers, Claude Opus 5.5 scores 1509 and ranks 3rd, Claude Fable 5.1 is 5th at 1504, and OpenAI's highest entry is GPT-5.6 Sol at 22nd with 1485; GPT-6 Astra sits 37th at 1476 (leaderboard updated Oct 27, 2026). Arena measures which answer people prefer across all kinds of prompts, not writing alone, so treat it as a strong hint rather than proof.

Hands-on tests point the same way. In Tom's Guide's Sep 24, 2026 test of ChatGPT-6 against Claude Opus 5.5, Claude won the rounds for messy planning and for writing that sounds human, and won four of five rounds overall. The same writer's Aug 27, 2026 test chose Claude as the writing coach over ChatGPT and Gemini. ChatGPT is not far behind: it won the vague-request round of the Sep 24 test, and a later Tom's Guide test found ChatGPT beat Anthropic's smaller Claude Haiku 5.5 by three rounds to two (Oct 15, 2026). These are one reviewer's prompts, not a benchmark.

Coding

Claude leads coding by a small margin. On Arena's WebDev board (updated Oct 27, 2026), Claude Opus 5.5 is 1st at 1810, GPT-6 Astra 2nd at 1773, Claude Sonnet 5.5 3rd at 1770 and GPT-6.1 Sol 5th at 1751. The Artificial Analysis Intelligence Index, which mixes coding, agent and reasoning tests, puts Claude Opus 5.5 at 59 against GPT-6 Astra's 57 (read Nov 6, 2026).

The makers' own tables disagree by test, as makers' tables do. Anthropic reports Opus 5.5 at 66.4% on Terminal-Bench 4.0 against GPT-6 Astra's 57.9%, while its same table shows GPT-6 Astra ahead on AutomationBench (41.4% to 40.0%) and Terminal-Bench-Science. Cost matters for heavy users: Artificial Analysis measured GPT-6 Astra at $6.24 per index task against $10.04 for Claude Opus 5.5.

The coding agents differ in who can use them. Codex is included on every ChatGPT plan, with limited access on Free and Go, while Claude Code needs Pro or higher. The agents themselves are compared on Codex vs Claude Code, and the developer platforms on OpenAI API vs Claude API.

Research and search

Both search the web on their free plans (ChatGPT, Claude). For long reports, ChatGPT's Deep Research is available, in limited form, even on Free, while Claude's Research feature is not included on Claude Free. On a free account, that makes ChatGPT the one to use for research.

On quality there is no reliable comparison yet. The only figures are the makers': Anthropic's table gives Claude Opus 5.5 67.7% on Humanity's Last Exam with tools against GPT-6 Astra's 57.2%, and OpenAI's own table also shows Claude ahead on that test, while OpenAI's shows GPT-6 Astra at 91.5% on BrowseComp, a web-browsing test, against Claude Opus 5's 90.8%. Neither tests the finished research reports people actually read.

Images and video

2:05 ChatGPT wins outright: it makes images and Claude does not. ChatGPT Images 2.6 is available on every plan, and since Nov 4, 2026 it also makes looping GIFs. Anthropic says Claude does not create photos or illustrations the way image tools do; it draws diagrams, charts and interactive visuals in HTML and SVG and can read images you upload. Neither makes video: OpenAI discontinued Sora on Apr 26, 2026.

Voice

ChatGPT's voice is the more finished product. Its Live voice is a regular feature with published allowances: 3 hours in a rolling 24 hours on Plus, 15 hours on Pro $100 and unlimited on Pro $200 (subject to abuse guardrails). Claude's voice mode is still labeled beta, though it runs on every Claude plan and can use connected tools such as Gmail and Calendar. No benchmark compares how the two sound or understand speech; one creator's experience is under What testers found below. More on ChatGPT's side: Voice Mode.

Price

At $20 a month the two are level: ChatGPT Plus and Claude Pro, and Claude Pro drops to $17 a month on a yearly plan, which ChatGPT does not offer for Plus. ChatGPT has more steps on the ladder: ChatGPT Go at $8 a month below Plus, and ChatGPT Pro at $100, $200 or $500 a month above it. Claude's step above Pro is Max at $100 and Ultra at $200. Both top tiers bundle API credits: Claude Max includes $100 a month and Ultra $200, and ChatGPT Pro includes $100, $200 or $500 depending on the tier. For teams, ChatGPT Business and Claude Team both start at $20 a seat a month billed yearly (Claude's too).

One difference money cannot fix on ChatGPT's cheap tiers: Free and Go may show ads, while Plus and above never do. See Ads in ChatGPT and ChatGPT Plans and Pricing.

Privacy

Claude has the friendlier default. OpenAI says data sharing is enabled by default on personal Free, Plus and Pro accounts, and you turn it off at Settings > Data controls > Improve the model for everyone while keeping your chat history. Anthropic says it uses consumer chats for training if you choose to allow it in the Model Improvement setting, and never uses incognito chats. Both may keep a conversation you rate with a thumbs up or down: Anthropic for up to 5 years, de-linked from your account. OpenAI does not train on ChatGPT Business data by default. ChatGPT's settings are on Privacy and Data Controls.

Limits

On free plans ChatGPT is far more generous: unlimited everyday text chats (subject to abuse safeguards, with separate limits on uploads, images and tools), while Claude Free has limited usage every 5 hours and every week. Neither publishes message counts for its $20 plan's chat. ChatGPT publishes ranges for agent work on Plus, for example 15 to 160 GPT-6.1 Sol messages every five hours in Work and Codex; Claude describes Max as 5x Pro usage and Ultra as having no 5-hour limit for verified accounts. Details: ChatGPT Usage Limits.

When to choose Claude, and when ChatGPT

Choose Claude for long writing, editing and coding, where the leaderboards and the hands-on tests of fall 2026 favor it, and if you want your chats kept out of training without changing a setting. Choose ChatGPT if you want images, a free plan you will rarely run out of, voice you can use for hours, a cheaper $8 plan, or a $500 plan with the most agent capacity. For other rivals, see When to Use Something Else.

Learn more: Claude on Ai.vi.ki and Claude vs ChatGPT on the Claude viki.

What testers found

1:50 In a Jun 8, 2026 video for The AI Advantage, Igor Pogany ran the same basic writing prompt in both and showed the results on screen. He called Claude's writing out of the box "less robotic" than ChatGPT's and gave it the writing round, while adding that both stop sounding robotic once you set them up with examples of your own writing. He gave the overall comparison to Claude, and images and voice to ChatGPT.

3:47 On voice, the same video found ChatGPT's voice input "close to perfect" and Claude's "flaky": in his use Claude misheard sentences and handled other languages, accents and punctuation worse. That is one person's experience of dictation, not a measured test.

10:54 Parker Prompts paid for the $20 plans of ChatGPT, Claude and Gemini for a month and, in an Aug 12, 2026 video, said Claude ran out of usage fastest by a clear margin and simply stops until the reset, while ChatGPT gave clear numbers, about 160 messages every 3 hours on its strongest model, and drops to a weaker model instead of locking you out.

4:39 Not every tester leans to Claude. In his Sep 2026 video on GPT-6 Astra's launch, Matt Wolfe said his own tests left him more impressed with Astra than with Claude Opus 5 or Fable 5, and on his own drawing-in-code test his AI judge placed GPT-6 first, ahead of Claude Fable 5.1.

Frequently asked questions

Is Claude better than ChatGPT? For writing and coding, Claude is slightly ahead on public leaderboards as of Nov 6, 2026: Claude Opus 5.5 ranks 3rd on Arena's text board and 1st on its web development board, against ChatGPT's best at 22nd and 2nd. ChatGPT is better for images, voice and free use.

Is Claude cheaper than ChatGPT? Not at the main tier: Claude Pro and ChatGPT Plus both cost $20 a month, though Claude Pro is $17 a month on a yearly plan. ChatGPT has a cheaper paid plan, Go at $8 a month, which Claude does not match.

Can Claude make images like ChatGPT? No. Claude draws diagrams, charts and interactive visuals in code, but it does not generate photos or illustrations; ChatGPT Images does on every ChatGPT plan, including Free.

Which is better for coding, ChatGPT or Claude? Claude Opus 5.5 leads Arena's WebDev leaderboard at 1810 against GPT-6 Astra's 1773 (Oct 27, 2026), a narrow lead. ChatGPT's Codex is available in limited form on the Free plan, while Claude Code needs Claude Pro or higher.

Does Claude train on my chats like ChatGPT? ChatGPT uses personal chats for training by default unless you turn off "Improve the model for everyone". Anthropic says it trains Claude on consumer chats only if you allow it in the Model Improvement setting.

Which has the better free plan, ChatGPT or Claude? ChatGPT Free, which gives unlimited everyday text chats plus limited images, deep research and Codex. Claude Free limits usage every 5 hours and every week and leaves out Claude Code and Research. ChatGPT Free may show ads in some countries.

Sources

ChatGPT pricing, OpenAI, read Nov 6, 2026
Pricing (Codex and ChatGPT plans), OpenAI, read Nov 6, 2026
ChatGPT Free Tier FAQ, OpenAI Help Center, read Nov 6, 2026
ChatGPT Voice, OpenAI Help Center, read Nov 6, 2026
Images in ChatGPT, OpenAI Help Center, read Nov 6, 2026
Searching the web with ChatGPT, OpenAI Help Center, read Nov 6, 2026
Data controls in ChatGPT, OpenAI Help Center, read Nov 6, 2026
Ads in ChatGPT, OpenAI Help Center, read Nov 6, 2026
Using Codex with your ChatGPT plan, OpenAI Help Center, read Nov 6, 2026
API credits benefit for eligible ChatGPT subscriptions, OpenAI Help Center, read Nov 6, 2026
What to know about the Sora discontinuation, OpenAI Help Center, read Nov 6, 2026
Introducing GPT-6 Astra, OpenAI, Sep 3, 2026
Plans and pricing, Anthropic, read Nov 6, 2026
Claude Opus 5.5, Anthropic, Sep 22, 2026
Can Claude produce images?, Anthropic Help Center, read Nov 6, 2026
Use voice mode, Anthropic Help Center, read Nov 6, 2026
Is my data used for model training?, Anthropic Privacy Center, read Nov 6, 2026
Text Arena leaderboard, Arena, updated Oct 27, 2026, read Nov 6, 2026
WebDev Arena leaderboard, Arena, updated Oct 27, 2026, read Nov 6, 2026
AI model leaderboards, Artificial Analysis, read Nov 6, 2026
I tested ChatGPT-6 vs Claude Opus 5.5 with 5 everyday prompts: it wasn't even close, Tom's Guide, Sep 24, 2026
I gave ChatGPT, Claude and Gemini the same draft: one was the best writing coach, Tom's Guide, Aug 27, 2026
I tested ChatGPT-6 vs Claude Haiku 5.5 on everyday tasks: here's the winner, Tom's Guide, Oct 15, 2026
ChatGPT vs. Claude: Which Should You Use in 2026?, The AI Advantage, Jun 8, 2026 (video)
ChatGPT vs Claude vs Gemini: Which $20 AI Plan Is Worth It?, Parker Prompts, Aug 12, 2026 (video)
GPT-6 Astra Is Finally Here (And It’s REALLY Good), Matt Wolfe, Sep 3, 2026 (video)
OpenAI announces it is shutting down video platform Sora, NBC News, Mar 24, 2026 (video)

Moments